Testing Procedure on QA system for SAP Roles?

[Former Member]

Hello All!

I am planning on designing a Test Process (SOP) on QA system:

As for now there is no UAT done on QA, we just create a Role on Security client(Dev) and import that to sandbox(we have data to test in this system) for testing the design. if everything fine we pust transport to QA(DEV-QA-PRD), we do role comparison sandbox vs QA(as part of QA testing) and push roles to PRD.

Occasionaly we bring business user for UAT (On sandbox).

I want to put a SOP on QA testing, any suggestions will be highly appreciated:

Other questions what I have is:

1) Is there any risk, if we create test users on QA system for testing the new roles?(I dont see any risk, as it is test system)

2)I want to bring business users to test their roles someplace, what will be the right system?

3)My managers one good reason why we need to testing agiain on QA while we do testing on sandbox and role comparison for QA testing? (when I say good practise, its hard for me to sell my idea).

I will appreciate all my team suggestions out there in SDN!

Note: We dont do testing on Dev clients as there is no data to support testing for the roles), our sandbox is a refresh of PRD so we do testing on sandbox.

Thanks All.

[Former Member]

Hi AJ,

I think you have answered your own questions

>1) Is there any risk, if we create test users on QA system for testing the new roles?(I dont see any risk, as it is test system)

No risk if they are restricted correctly (like using the production access).

>2)I want to bring business users to test their roles someplace, what will be the right system?

I would use your QA system which has correct configuration and data

>3)My managers one good reason why we need to testing agiain on QA while we do testing on sandbox and role comparison for QA testing? (when I say good practise, its hard for me to sell my idea).

Authorisation checks are dependent on config and data. If you cannot test with realistic configuration and data then you cannot say that the test is completed correctly. Generally Sandbox data and config is not "production like" so the testing has limited validity. *note I just re-read about the refresh from Prod, if this is the case then the risk of differences is reduced a bit, however if QA is more up-to-date then this is what you should use.

Edited by: Alex Ayers on May 21, 2009 3:18 PM

[Former Member]

Thanks Alex for quick response! I am with you on the suggestion.

Like I said QA and Sandbox are almost refreshed at same time with PRD, so the master or config data what we have on no productive clients are same.

I did not role comparison tool failing anytime, did any any in our forum saw any discrepancies with role comparison? If there isn't any I think I have to stick to regular old procedure, as I cannot sell my idea to management.

Again I really appreciate your suggestion! i will wait if my team have any more suggestions for me!

[Former Member]

Great answer by Alex and he has given you more info,only thinh i can add is testing without actual data(in sandbox) is very easy as it has very limited data compared to data you have in production,so in QA if you have client refresh acitvity and you get more data to test the execution ,you testing is more appropriate compared to sandboz,what Alex said is perfect

Thanks,

Prasant K Paichha